The goal of cancer research for the past 50 years has been to identify drugs that kill cancer cells while leaving normal tissues unharmed. To a large extent that search has been unsuccessful: currently used chemotherapeutic agents are general cytotoxins that are useful only because they exhibit a marginal degree of selectivity in killing cancer cells. The reasons for even this meager selectivity are unknown although processes like apoptosis, cell cycle checkpoints and DNA damage repair probably play a role. It is also likely that taken together, the factors that are responsible for that selectivity create a cellular environment or context, that could be exploited by a drug to a greater degree. A promising new anticancer drug, the depsipeptide FR-901,228 has exhibited this kind of context-specific anticancer activity in preliminary studies. The following proposal uses the example of FR-901,228 to outline a paradigm for anticancer chemotherapy that relies on context-specific toxicity. FR-901,228 is a structurally-unique bacterial metabolite that was identified because of its ability to change the morphology of ras-transformed cells to that of normal cells. Its mechanism of action is not known. By extending our preliminary work, we will identify the molecular target of the drug and its mechanism of action in coordinated studies that combine synthetic chemistry, cell biology and genetics. We will also use cell-based studies to identify the factors that determine how cells respond the drug. Specific Aim 1 is to synthesize a biologically active analog that will be used to prepare an FR-901,228 affinity matrix and a radiolabeled drug. This will be done by adapting the strategy we used for the first synthesis of FR-901,228 itself. Specific Aim 2 is to use the biological activity of FR-901,228 to determine the cellular factors that influence the drug's toxicity. We will determine the sensitivity of matched pairs of cell lines containing cancer-related defects to the drug and correlate it with specific alterations found in tumors. Specific Aim 3 is to identify the drug's target in a model system through genetic means. We have shown that S. pombe growth is inhibited by FR-901,228 and we will use this as the basis to identify genes that, when overexpressed, render S. pombe resistant to the drug. We will also use this selection strategy to identify drug-resistant mutants. Together, the studies outlined in our Specific Aims, will constitute the first step in testing the hypothesis of context-specific toxicity in cancer chemotherapy.